Sister pair DE heatmap ctrl poly

library(Seurat)

Attaching SeuratObject

library(dplyr)

Attaching package: ‘dplyr’

The following objects are masked from ‘package:stats’:

    filter, lag

The following objects are masked from ‘package:base’:

    intersect, setdiff, setequal, union

library(ggplot2)
library(stringr)
library(tibble)
library(patchwork)
library(plotly)

Attaching package: ‘plotly’

The following object is masked from ‘package:ggplot2’:

    last_plot

The following object is masked from ‘package:stats’:

    filter

The following object is masked from ‘package:graphics’:

    layout

DE table

First we load the sister pair DE tables and filter for:

• absolute avg_log2FC > 0.5 (~41% increase)

• p_val_adj < 0.01

DE_list <- readRDS("~/spinal_cord_paper/data/Gg_ctrl_poly_sis_markers.rds")

for (i in seq(DE_list)) {
    DE_list[[i]] <- DE_list[[i]] %>% 
    arrange(desc(avg_log2FC)) %>% 
    filter(abs(avg_log2FC) > 0.5) %>% 
    filter(p_val_adj < 0.01)
}

DE_table <- do.call(rbind, DE_list)
dim(DE_table)

[1] 807   8

delta pct distribution

hist(abs(DE_list[[1]]$delta_pct), breaks = 20)
abline(v = 0.1, lty = "dashed", col = "red")

hist(abs(DE_list[[2]]$delta_pct), breaks = 20)
abline(v = 0.1, lty = "dashed", col = "red")

hist(abs(DE_list[[4]]$delta_pct), breaks = 20)
abline(v = 0.1, lty = "dashed", col = "red")

hist(abs(DE_list[[5]]$delta_pct), breaks = 20)
abline(v = 0.1, lty = "dashed", col = "red")

Now we filter the DE lists for absolute delta percentage > 0.1.

for (i in seq(DE_list)) {
  DE_list[[i]] <- DE_list[[i]] %>% 
  filter(abs(delta_pct) > 0.1)
}

DE_table <- do.call(rbind, DE_list)
dim(DE_table)

[1] 671   8

Broad clusters

broad_order <- c("progenitors",
      "FP",
      "RP",
      "FP/RP",
      "neurons",
      "OPC",
      "MFOL",
      "pericytes",
      "microglia",
      "blood",
      "vasculature"
      )

Integrated data

Load the integrated control and poly data.

int_path <- "Gg_ctrl_poly_int_seurat_250723"

my.se <- readRDS(paste0("~/spinal_cord_paper/data/", int_path, ".rds"))
  annot_int <- read.csv(list.files("~/spinal_cord_paper/annotations",
                               pattern = str_remove(int_path, "_seurat_\\d{6}"),
                               full.names = TRUE))
  
  if(length(table(annot_int$number)) != length(table(my.se$seurat_clusters))) {
     stop("Number of clusters must be identical!")
  }
  
  # rename for left join
  annot_int <- annot_int %>% 
    mutate(fine = paste(fine, number, sep = "_")) %>% 
    mutate(number = factor(number, levels = 1:nrow(annot_int))) %>% 
    rename(seurat_clusters = number)
  
  ord_levels <- annot_int$fine[order(match(annot_int$broad, broad_order))]
   
  # add cluster annotation to meta data
  my.se@meta.data <- my.se@meta.data %>% 
    rownames_to_column("rowname") %>% 
    left_join(annot_int, by = "seurat_clusters") %>% 
    mutate(fine = factor(fine, levels = ord_levels)) %>% 
    mutate(seurat_clusters = factor(seurat_clusters, levels = str_extract(ord_levels, "\\d{1,2}$"))) %>% 
    column_to_rownames("rowname")
  
  ctrl_poly_int_combined_labels <- readRDS("~/spinal_cord_paper/annotations/ctrl_poly_int_combined_labels.rds")
  
  my.se <- AddMetaData(my.se, ctrl_poly_int_combined_labels)
  

DimPlot

DimPlot(
  my.se,
  group.by = "annot_sample",
  reduction = "tsne",
  label = TRUE,
  repel = TRUE
  ) +
  NoLegend()

Cluster order

Get the cluster order from the spearman correlation heatmap of the control and poly integrated data. Then we filter for the neuronal clusters only.

corr_heatmap <- readRDS("~/spinal_cord_paper/output/heatmap_spearman_ctrl_poly.rds")

#heatmap order
htmp_order <- data.frame("label" = corr_heatmap[["gtable"]]$grobs[[4]]$label) %>% 
  mutate(label = str_remove(label, "_int")) %>% 
  mutate(label_ordered = paste(str_sub(label,6 ,-1), str_sub(label, 1, 4), sep = "_"))

my.se@meta.data <- my.se@meta.data %>%
  mutate(annot_sample = factor(annot_sample, levels = htmp_order$label_ordered))

Idents(my.se) <- "annot_sample"

# filter for the neuronal clusters
my.se <- subset(my.se, idents = htmp_order$label_ordered[grepl("neurons|MN|CSF", htmp_order$label_ordered)])

DimPlot(
  my.se,
  group.by = "annot_sample",
  reduction = "tsne",
  label = TRUE,
  repel = TRUE
  ) +
  NoLegend()

my.se@active.assay <- "RNA"

Dotplot

# Dotplot of sister pair makrers
pl_all <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample", 
                      # reverse order of unique genes so number one is on top
                    features = rev(unique(DE_table$Gene.stable.ID)),
                    gnames = modplots::gnames,
          cols = c("lightgrey", "black")) +
    theme(axis.text.x = element_text(angle = 90, hjust=1, vjust=0.5)) +
    coord_flip()

pl_all

pdf("~/spinal_cord_paper/figures/Sister_pair_DE_dotplot.pdf", width = 15, height = 32)

  pl_all

  
DE_table$Gene.name[duplicated(DE_table$Gene.stable.ID)]

 [1] "HES5"               "MAP6"               "GNG5"               "ST18"               "GAD1"               "FABP3"             
 [7] "SYT1"               "SLC32A1"            "KIF5C"              "HMP19"              "GALNT9"             "VSTM2L"            
[13] "HINTW"              "DNER"               "CRABP-I"            "RELN"               "PAX2"               "NEUROD2"           
[19] "CHL1"               "LHX1"               "NRXN3"              "ENSGALG00000029521" "BHLHE22"            "SPOCK1"            
[25] "SSTR1"              "SLC32A1"            "NCALD"              "ID2"                "GRIK3"              "GAD2"              
[31] "PTPRK"              "GABRG3"             "GAD1"               "RUNX1T1"            "HPCAL1"             "ZEB2"              
[37] "GALNT9"             "ENSGALG00000013212" "MDK"                "ZFPM2"              "RELN"               "NEUROD6"           
[43] "CPLX1"              "LAMP5"              "WNT5A"              "HINTW"              "SOX4"               "DKK3"              
[49] "UNC13B"             "ATP1B1"             "GALNT17"            "RASD1"              "ENSGALG00000051980" "PLXNA4"            
[55] "DACT2"              "DISP3"              "MVB12B"             "ENSGALG00000054223" "CNTN4"              "ZNF423"            
[61] "CBLB"               "FKBP1B"             "CELF2"              "EPB41L4A"           "PXYLP1"             "ENSGALG00000023640"
[67] "CNTN2"              "MRPS6"              "PPP3CA"             "NFIX"               "NFIA"               "SOX8"              
[73] "DRAXIN"             "CRABP-I"            "NHLH1"              "TAC1"               "VSTM2L"             "CPNE2"             
[79] "PRKCA"             

Individual dot plots

# select top50 by log2FC 
for (i in seq(DE_list)) {
    DE_list[[i]] <- DE_list[[i]] %>%
    slice_max(order_by = abs(avg_log2FC), n = 50) %>% 
    arrange(desc(avg_log2FC))
}

p1 <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample", 
                    assay = "RNA",
                      # reverse order of DE genes so number one is on top
                    features = rev(DE_list[[1]]$Gene.stable.ID),
                    gnames = modplots::gnames,
          cols = c("lightgrey", "black")) +
    theme(axis.text.x = element_blank()) +
    coord_flip() +
    xlab(names(DE_list)[1]) +
    ylab(element_blank())

p2 <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample",  
                    assay = "RNA",
                      # reverse order of DE genes so number one is on top
                    features = rev(DE_list[[2]]$Gene.stable.ID),
                    gnames = modplots::gnames,
          cols = c("lightgrey", "black")) +
    theme(axis.text.x = element_blank()) +
    coord_flip() +
    xlab(names(DE_list)[2]) +
    ylab(element_blank())

p3 <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample",  
                    assay = "RNA",
                      # reverse order of DE genes so number one is on top
                    features = rev(DE_list[[5]]$Gene.stable.ID),
                    gnames = modplots::gnames,
          cols = c("lightgrey", "black")) +
    theme(axis.text.x = element_blank()) +
    coord_flip() +
    xlab(names(DE_list)[5]) +
    ylab(element_blank())

p4 <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample",  
                    assay = "RNA",
                      # reverse order of DE genes so number one is on top
                    features = rev(DE_list[[4]]$Gene.stable.ID),
                    gnames = modplots::gnames,
          cols = c("lightgrey", "black")) +
    theme(axis.text.x = element_text(angle = 90, hjust=1, vjust=0.5)) +
    coord_flip() +
    xlab(names(DE_list)[4]) +
    ylab(element_blank())

layout <- "CCDD
           CC##"

pdf("~/spinal_cord_paper/figures/Supp_Fig_5_ctrl_poly_dotplot_individual.pdf", height = 21, width = 7)
# without labels for proper alignment
(p1 + p2 + plot_layout(guides = "collect")) /
(p3 + p4 + plot_layout(guides = "collect", design = layout)) & 
  theme(axis.text.x = element_blank(),
        axis.text.y = element_blank())
# with labels to transfer in illustrator
(p1 + p2 + plot_layout(guides = "collect")) /
(p3 + p4 + plot_layout(guides = "collect", design = layout))

dev.off()

null device 
          1 

Volcanoplots

p.adj <- 0.01
l2fc <- 0

# select top50 by log2FC 
for (i in seq(DE_list)) {
    DE_list[[i]] <- DE_list[[i]] %>% 
    mutate(delta_pct_sign = case_when(
      delta_pct < 0 ~ "-",
      delta_pct > 0 ~ "+",
      delta_pct == 0 ~ "0"
    ))
}
 

c16_p14 <- ggplot(data = DE_list[[4]],
       aes(x = avg_log2FC,
           y = -log10(p_val_adj),
           label = Gene.name,
           color = delta_pct_sign,
           size = abs(delta_pct)/2
       )) +
  geom_point(shape = 21) +
  geom_hline(yintercept = -log10(p.adj), linetype = "dashed") +
  geom_vline(xintercept = c(-l2fc,l2fc), linetype = "dashed") +
  scale_color_manual(values = c("red", "black")) +
  scale_size_continuous(range = c(0.5, 4)) +
  xlim(-2,2) +
  ylab("-log10(padj)") +
  theme_bw()

c11_p15 <- ggplot(data = DE_list[[1]],
       aes(x = avg_log2FC,
           y = -log10(p_val_adj),
           label = Gene.name,
           color = delta_pct_sign,
           size = abs(delta_pct)/2
       )) +
  geom_point(shape = 21) +
  geom_hline(yintercept = -log10(p.adj), linetype = "dashed") +
  geom_vline(xintercept = c(-l2fc,l2fc), linetype = "dashed") +
  scale_color_manual(values = c("red", "black")) +
  scale_size_continuous(range = c(0.5, 4)) +
  xlim(-2,2) +
  ylab("-log10(padj)") +
  theme_bw()

ggplotly(c16_p14)

ggplotly(c11_p15)

NA

pdf("~/spinal_cord_paper/figures/Fig_5_volcanoplots.pdf", width = 7, height = 10)
(c16_p14 +
  ggrepel::geom_text_repel(size = 3, color = "black")) /
(c11_p15  +
  ggrepel::geom_text_repel(size = 3, color = "black")) + plot_layout(guides = "collect")

Warning: Removed 1 rows containing missing values (geom_point).
Warning: Removed 1 rows containing missing values (geom_text_repel).
Warning: ggrepel: 10 unlabeled data points (too many overlaps). Consider increasing max.overlaps

Specific markers

Find Markers for clusters 11_ctrl, 16_ctrl, and 15_poly.

gnames <- modplots::gnames

markers <- list()

clu <- c("inhibitory_neurons_16_ctrl",
         "excitatory neurons_11_ctrl",
         "excitatory_neurons_15_poly")

for (i in seq(clu)) {  
  markers[[i]] <- FindMarkers(
      my.se,
      ident.1 = clu[i],
      group.by = "annot_sample",
      assay = "RNA",
      verbose = FALSE,
      only.pos = TRUE, # we look for overexpressed, specific markers
      min.pct = 0.25,
      logfc.threshold =  0.25,
      latent.vars = c("CC.Difference.seurat"),
      test.use = "MAST"
    ) %>%
      tibble::rownames_to_column("Gene.stable.ID") %>%
      dplyr::left_join(gnames, by = "Gene.stable.ID") %>%
      dplyr::arrange(-avg_log2FC) %>%
      dplyr::filter(p_val_adj < 0.05) %>%
      dplyr::filter(abs(avg_log2FC) > 0.5) %>%
    dplyr::mutate(delta_pct = abs(pct.1 - pct.2))
}

names(markers) <- clu

Specific marker dotplot

Plot the top 50 markers for clusters 11_ctrl, 16_ctrl, and 15_poly.

n <- 50

mark_plot <- list()

for (i in seq(clu)) {
  mark_plot[[i]] <- modplots::mDotPlot2(my.se,
                    group.by = "annot_sample", 
                      # reverse order of markers so number one is on top
                    features = rev(markers[[i]][1:n,"Gene.stable.ID"]),
                    gnames = modplots::gnames) +
    theme(axis.text.x = element_text(angle = 90, hjust=1, vjust=0.5)) +
    coord_flip() +
  scale_colour_gradientn(colours = c("gray90","gray80","yellow", "orange", "red", "darkred", "darkred")) +
  ggtitle(paste0("Top ", n, " markers by log2FC for ", clu[i]))

}

mark_plot[[1]]
mark_plot[[2]]
mark_plot[[3]]


pdf("~/spinal_cord_paper/figures/Sister_pair_neuron_marker_dotplots.pdf", width = 14, height = n/3)
mark_plot[[1]]
mark_plot[[2]]
mark_plot[[3]]

# Date and time of Rendering
Sys.time()

sessionInfo()